Development of fundamental understanding, characterisation and modelling techniques for soft magnetic materials in industrial applications

Washbourne, Alexander 2025. Development of fundamental understanding, characterisation and modelling techniques for soft magnetic materials in industrial applications. EngD Thesis, Cardiff University. Item availability restricted.

Preview	PDF (Thesis) - Accepted Post-Print Version Download (21MB) | Preview
	PDF (Cardiff University Electronic Publication Form) Restricted to Repository staff only Download (358kB)

Abstract

This thesis advances the understanding and modelling of non-sinusoidal losses in non-oriented electrical steels (NOES) subjected to harmonic-rich and inverterdriven excitations. The research addresses a critical industrial challenge: how to accurately predict and mitigate core losses to optimise material usage and machine efficiency. The study pursued four objectives, including direct measurement of harmonic induced losses, development of harmonically rich testing protocols, model refinement, and empirical validation. Experimental validation of quasi-static and dynamic test systems confirmed their reliability for characterising hysteresis and eddy current losses. A comprehensive dataset capturing NOES behaviour under single, combined, and PWM harmonic excitations was established, revealing the limited benefits of ultra-thin laminations (0.1 mm) compared to 0.12 mm grades under harmonic-rich conditions due to hysteresis saturation. A key modelling innovation introduced a frequency-dependent permeability correction and dynamic scaling factor to refine loss predictions. This approach, validated against empirical data, outperformed cumulative models and aligns with a harmonic-resolved CLS framework capable of accounting for the dominance of dynamic and excess losses beyond the fundamental frequency. The outcomes of this research provide industry stakeholders with improved predictive tools for designing efficient, inverter-fed magnetic cores, contributing to advances in traction motors, wind turbines, and high-speed industrial drives. Future work will focus on extending the dataset to higher frequencies, mitigating hysteresis saturation, and embedding adaptive modelling techniques, including machine learning, for real-time predictive capability.

Item Type:	Thesis (EngD)
Status:	Unpublished
Schools:	Schools > Engineering
Uncontrolled Keywords:	1. Non-oriented electrical steel 2. Classical loss separation 3. Relative permeability 4. Skin depth 5. Frequency dependent modelling 6. Energy efficiency
Date of First Compliant Deposit:	3 July 2025
Last Modified:	03 Jul 2025 15:23
URI:	https://orca.cardiff.ac.uk/id/eprint/179524

Actions (repository staff only)

Edit Item